IF FACTOR SHARES ARE NOT CONSTANT THEN WE HAVE A MEASURMENT PROBLEM. CAN WE SOLVE IT? Hernando Zuleta SERIE DOCUMENTOS DE TRABAJO No. 67 Agosto 2009 IF FACTOR SHARES ARE NOT CONSTANT THEN WE HAVE A MEASURMENT PROBLEM. CAN WE SOLVE IT? 1 Hernando Zuleta Universidad del Rosario American University in Bulgaria [email protected] Abstract Recent evidence show that factor shares, if properly measured, are far from constant. Moreover, the shares of natural resources and raw labor seem to be negatively correlated with income per capita while the share of human and physical capital is positively correlated with income per capita. Now, if factor shares are not constant then (i) growth accounting exercises rely on a false assumption and (ii) there is a measurement problem. The effect that changes in factor shares have on output depend on the relative abundance of factors and, for this reason, it is necessary to have correct measures. We propose an empirical methodology to solve the measurement issue and estimate TFP growth. JEL Codes: O11, O30, O41 Keyword: Factor Shares, Production Function, Measurement. 1 The findings, recommendations, interpretations and conclusions expressed in this paper are those of the authors and not necessarily reflect the view of the Department of Economics of the Universidad del Rosario. 1 1. Introduction The idea that labor income share is roughly constant, namely, the Cobb-Douglas-Kaldor paradigm2, produced important consequences in the area of economic growth. Almost all of the literature on growth accounting assumes that the elasticity of output with respect to capital is constant and concludes that the major part of economic growth is explained by growth in TFP1 (see Easterly and Levine, 2002; Solow, 1957 or Young, 1994 among others). Recent empirical work shows that the Cobb-Douglas-Kaldor paradigm is not really supported by the data. Kahn and Lim (1998) show that the shares of equipment, production workers and non production workers have clear trends. Blanchard (1997) observes the share of labor decreases in continental Europe after the 80s and suggests that the reason of such decline may be technological bias. Other authors calculate the income share of reproducible factors (human and physical capital) and not reproducible factors and it turns out that the later is positively correlated with the income level (see Krueger, 1999; Caselli and Feyrer, 2007; Zuleta, 2008a; Sturgill, 2009 and Zuleta, Parada and Campo, 2008)3. Additionally, the variability of factor shares may have different effects on output depending on the factor abundance of the economy. If the income share of abundant factors is growing the effect of these changes on the income level is positive but if income share of abundant factors is decreasing the effect of these changes on the income level is negative. In this paper we propose an empirical methodology to estimate the correct measures of factor shares. In the next section we explain why measurement is an issue when factor shares are not constant. In the third section we present the methodology. In the last section we present the conclusions. 2 See Cobb and Douglas (1928) and Kaldor (1961). In the last decades some authors revisited the theory of biased innovations and challenged the CobbDouglas-Kaldor paradigm. Zeira (1998) provides a model of technological innovations that reduce labor requirements and find that innovations are only adopted in countries with high wages. Seater (2005), Zeira (2006), Peretto y Seater (2007) and Zuleta (2008b) among others, present models where the scarcity of not reproducible factors generates incentives to adopt technologies that reduce the need of the these factors. In a market economy, this type of technological change affects factor income shares in such a way that the share or reproducible factors is positively correlated to the income level of the economy. 3 2 2. Factor Shares and Measurement To illustrate the importance of correct measures for the factors consider the simplest CobbDouglas technology with two factors: K and L. Output per worker, y as a function of capital per worker k Y , can be expressed L K : L y Ak Now, suppose that there is an increase in the share of capital. The effect of the change in factor shares on income per worker depends on the relative abundance of capital, y Ak ln k Therefore, if k 1 the effect is positive and if k 1 the effect is negative. Figure 1 illustrates this fact. Figure 1 y 1 k 3. How to solve the problem? For simplicity consider a production function with only two factors: (1) Yt At K K t t L Lt 1 t 3 Where Y is total income, K is physical capital L is raw labor. Note that factor shares have the subindex t, namely they are variables not parameters. Now, given that factor shares vary, the relative abundance of factors becomes very important and, for this reason, it is necessary to have correct measures of capital and labor. The parameters K and L play this role. If K K t 1 then the economy is capital abundant. L Lt a. The Methodology assuming away factor augmenting technological change. Differentiating equation 1, K Yt At K t L t (1 t ) t t ln K t Yt Kt Lt L Lt At (2) Now define (3) St Yt Kt L t (1 t ) t Yt Kt Lt The variable S in equation 3 is the Solow residual. Now, from equations 2 and 3 it follows that the Solow residual is not only TFP growth. The Solow residual includes biased technological change: (4) St K At t ln K t At L Lt K ~ Define St St t ln t Lt So equation 4 can be rewritten (5) ~ A St t t ln K At L Therefore, the reduced form to be estimated is the following, ~ St C0 C1 t t (6) Where At C0 t , C1 ln K At L . 4 Therefore, this methodology allows us to identify the correct measures of factors per worker K . L Finally, combining the previous results with equation 1 and taking logs we get K log Yt log At t log K t log(L Lt ) L Lt (7) We also know that log At log A0 log(1 C0 t ) . Therefore, we can define K ~ Yt log Yt t log K t log(1 C0 t ) log Lt so that equation 7 can be written L Lt ~ as Yt log( A0L ) . Therefore, we can also identify the initial level of TFP multiplied by the correct measure of labor. b. What if there is labor augmenting or capital augmenting technological change? We assume that there is factor augmenting technological change and the rate of technological change is constant. Equation 2 becomes AK t AL t At A t A (1 t ) A Lt Kt Yt t Yt K t (1 ) Lt ln K , t K t t t t L Kt Lt L ,t t (2A) Where AL t AK t and are, respectively, capital and labor augmenting technological AK t AL t change. Equations 4 and 5 becomes (4A) A K A A St t t K t (1 t ) L t t ln K ,t t AK t AL t At L ,t Lt (5A) A A ~ St t ln K , 0 1 K t L t L , 0 AK t AL t AK t AL t At AL t t A AL t AL t t AK t 5 Therefore, the reduced form to be estimated is the following, (6) ~ St C0 C1 t C2 t t Where C1 ln K L AK t AL t 1 AK t AL t A A , C2 K t L t A AL t Kt A A and C0 t t L t . At AL t Therefore, we can identify the correct measures of factors per worker K , the difference L between capital augmenting and labor augmenting technological change and the sum of neutral plus labor augmenting technological change AK t AL t AK t AL t At AL t . At AL t References Acemoglu, Daron (2002) "Directed Technical Change" the Review of Economic Studies, 69 (4) 781-809. Boldrin, Michele and Levine, David K. (2002) " Factor Saving Innovation" Journal of Economic Theory, 105 (1), 18-41 Caselli, Francesco and Feyrer, James (2007) "The Marginal Product of Capital" Quarterly Journal of Economics, vol. 122(2), 535-568, 05. Cobb C W and Douglas P H (1928) "A Theory of Production", American Economic Review, 18 (Supplement), 139-165 Dunning, J.H. (1998), Explaining International Production London: Unwin Hyman. Gollin, D., (2002) “Getting Income Shares Right”, Journal of Political Economy, 110 (2), 458-474. Krueger, Alan (1999) "Measuring Labor's Share" American Economic Review, 89 (2), 45-21. Heckscher, Eli F., (1919) “The Effect of Foreign Trade on the Distribution of Income”, Ekonomisk Tidskrift Kaldor, N., (1961), Capital Accumulation and Economic Growth, In FA Lutz and DC Hague, eds., The Theory of Capital, 177-222. New York St, Martin’s Press. Kennedy, Charles (1964), "Induced Bias in Innovation and the Theory of Distribution", Economic Journal, LXXIV, 541-547. Peretto, Pietro and Seater, John "Augmentation or Elimination?", working paper 2006, http://ideas.repec.org/p/deg/conpap/c011_060.html Sturgill, Brad (2009) “Cross-country Variation in Factor Shares and its Implications for Development Accounting” working Papers 09-07, Department of Economics, Appalachian State University. Young, A., (1995). The Tyranny of Numbers: Confronting the Statistical Realities of the East Asian Growth Experience. Quarterly Journal of Economics, 110: 641-680 Zeira, Joseph. (1998) "Workers, Machines and Economic Growth." Quarterly Journal of Economics, 113 (4), pp. 1091-1117. Zeira, Joseph, (2005) "Machines as Engines of Growth," CEPR Discussion Papers 5429. 6 Zuleta, Hernando, (2007) “Why labor income shares seem to be constant,” Journal of International Trade & Economic Development, 16 (4), 551-557. Zuleta, Hernando, (2008a) “An Empirical Note on Factor Shares,” Journal of International Trade & Economic Development, 17 (3), 379-390. Zuleta, Hernando, (2008b) “Factor Saving Innovations and Factor Income Shares,” Review of Economic Dynamics, 11 (4), 836-851. Zuleta, Hernando; Parada, Julian y Campo, Jacobo (2008) "Capital Natural, Capital Humano y Participación de los Factores. Una Revisión de los Métodos de Medición del Crecimiento Económico," DOCUMENTOS DE TRABAJO 004713, UNIVERSIDAD DEL ROSARIO - FACULTAD DE ECONOMÍA. 7